From the Hurricane Georges disaster that struck Puerto Rico, a database with unique potential to address many of the methodological short-comings of much of previous child disaster mental health research has emerged. This two-wave epidemiological study of child and adolescent psychiatric disorders improves upon many prior studies on children's post-disaster mental health and recovery through its sampling strategy, high quality measurement of psychiatric disorders, breadth and depth of measurement of socio-ecological risk and protective factors, and the use of multiple informants. The aims of this secondary data analysis study are examining how the recovery environment, within the domains of the family, peer, school, and community microsystems, influences the relationship between disaster exposure and the development of an internalizing disorder at 18 and 30 months post-disaster. Specific hypotheses about the relative influence of constructs within each microsystem and their mechanism of influence are addressed. Within the family microsystem, this includes assessing aspects of family relationship quality over time, its roles as a mediator in the exposure-disorder relationship, and potential moderators of family relationship quality. We will examine how peer relationships influence risk of psychiatric disorder following a disaster, with a specific focus on the influence of deviant peers and how peer victimization may erode peer support. In relation to the school microsystem, we assess the influence of school violence and efficacy. We also address how living within communities characterized by high levels of violence and poor neighborhood quality influence recovery. Children aged 4-17 years were randomly-selected from an island-wide probability household sample. Data were weighted to represent the child and adolescent population in Puerto Rico as indicated by the 2000 U.S. Census. Structured interviews using the Diagnostic Interview Schedule for Children and measures of risk and protection were conducted using highly trained interviewers at 18 and 30 months post-disaster. Interviews were completed with 1,886 parent-child dyads representing a completion rate of 90.1 percent. At time 2, there was a 94 percent retention rate. Multi-level structural equation modeling will be used to address the aims of the study. This thorough investigation of how risk and protection within the post-disaster recovery environment influences the development of an internalizing disorder will help inform and target multi-level interventions to reduce risk for adverse mental health outcomes following a disaster. PUBLIC HEALTH RELEVANCE: This study examines children's mental health in the long-term aftermath of a disaster. Child, family, and school/community factors that increase or decrease risk for developing a psychiatric disorder following exposure to a disaster are studied.